Seventeen limits that have to be asked for

Seventeen places have a limit on staying that nothing in the room says. It exists, and it is produced on request.

Eight per cent of the record, and the least knowable thing in it.

Where they are

cafés and coffee shops               10
pubs and bars                         3
shopping centres and arcades          2
laundrettes and shops with chairs     1
community and church halls            1
                                     ---
                                      17

Thirteen of the seventeen are cafés and pubs: the two kinds that are never free at the door and that between them make up forty-three per cent of the record.

What they say

The median told limit is ninety minutes. The written median is two hours.

So a limit that has to be asked for is, on average, half an hour shorter than one a place has printed. A team converting this walk into repeatable fieldwork can review how the platform handles this before choosing which time and project records are genuinely necessary.

Thirty-eight observations split seventeen and twenty-one. The difference is real in this record and it is not defensible as a finding about anywhere, and the two readings that fit it are incompatible. For wider research on public space, access and how places are used, consult American Planning Association.

It is a genuine rule. Some places have a shorter policy and have not bothered with a notice, because a notice would be unfriendly and the situation hardly ever arises.

It is an answer to an awkward question. Somebody is asked, in their own room, how long a stranger may sit. Ninety minutes is a plausible, safe, slightly defensive thing to say, and it may not describe what the place does.

The record cannot separate these and does not try. Every page using the seventeen says which it is relying on, and none of them relies on the ninety minutes being a policy.

What asking costs the record

A question changes a room.

The instrument was one sentence to one member of staff, asked only where no notice was found. That is as light as the question can be made, and it is still a stranger asking somebody at work to improvise a rule about a situation they may never have considered.

Seventeen of these limits may have come into existence at the moment they were asked for.

That possibility cannot be ruled out from inside the record, and it is the single largest doubt about the limit field. If all seventeen were inventions, the real figure for places with a limit would be twenty-one — ten per cent — and the headline would be stronger, not weaker.

So the doubt runs in the direction that would flatter this site, which is the reason it is stated at length rather than in a footnote.

Against the assumption

Of the seventeen, eight were assumed to have a limit and nine were not.

That is the worst anticipation rate in the record: less than half. A written limit was anticipated fifteen times out of twenty-one; a told one, eight times out of seventeen.

The reason is not mysterious. A notice is visible and a policy in somebody's head is not. What is worth noticing is that the invisible rules are also the shorter ones, so a person relying on their own judgement in a café is wrong in both directions at once: they assume a limit where there is none, and they miss the shorter ones where there is.

The ten cafés

More than half the told limits are cafés, and they are the same cafés that produce most of the rest of this record's awkwardness.

Of fifty-eight cafés: nineteen have a limit, of which ten must be asked for. Thirty-nine have none. Forty-seven were assumed to have one, and thirty-four of those assumptions were wrong.

A café is the place where a stranger is least able to find out where they stand by looking, most likely to guess, and most likely to guess wrong.

What a told limit is not

Not a policy document. Nobody saw one. The answer came from whoever was behind the counter.

Not necessarily consistent. A different member of staff on a different day might have said something else, and in two places two members of staff were present and only one was asked.

Not a refusal. Nobody in the walk was asked to leave, and no answer was given unprompted. All seventeen are responses to a question.

The seventeen by kind, against the written twenty-one

kind              told  written  total limits  of
cafés              10      9        19         58
pubs                3      1         4         31
shopping centres    2      4         6         16
laundrettes         1      2         3         12
community halls     1      1         2         11
libraries           0      2         2         12
covered seating     0      2         2         12

Two kinds print more than they say: shopping centres and libraries, both of them free at the door, both of them places with a wall and no till.

Two kinds say more than they print: cafés and pubs, both of them businesses with somebody behind a counter.

That is the cleanest version of the pattern. Where there is a person to ask, the rule lives in the person. Where there is not, it goes on the wall.

What this does to the record's firmest figure

The hundred and sixty-eight places with no limit rest on a look and one question.

In twenty-one of the hundred and eighty-five places where a question was needed, there was nobody obvious to ask, and the look stood alone. All twenty-one are in the no-limit column.

If those twenty-one behave like the rest — where roughly nine per cent of asked places produced a told limit — then perhaps two of them have a limit nobody could report. That would move the figure from a hundred and sixty-eight to a hundred and sixty-six.

The correction is small, it is stated, and it runs against this site's headline rather than for it.

Why seventeen is reported at all

Because dropping them would have made the record's own figures look better.

A limit field with only written notices in it would read: twenty-one places in a city limit staying, ten per cent, almost nowhere. That is a cleaner sentence and it leaves out seventeen places that, asked directly, said there was a limit.

The method took the answer it got. Where the answer is soft, the pages say so.